fungal genetics and biologyreal.mtak.hu/21277/1/selection against somatic parasitism...1 3 selection...

1

3

4

5

6

7 Q1

8 Q29

1011

1 3

14151617

1819202122232425

2 6

50

51

52

53

54

55

56

57

58

59

60

61

62

Q3

Fungal Genetics and Biology xxx (2014) xxx–xxx

YFGBI 2736 No. of Pages 10, Model 5G

17 October 2014

Contents lists available at ScienceDirect

Fungal Genetics and Biology

journal homepage: www.elsevier .com/locate /yfgbi

Selection against somatic parasitism can maintain allorecognitionin fungi

http://dx.doi.org/10.1016/j.fgb.2014.09.0101087-1845/� 2014 Published by Elsevier Inc.

⇑ Corresponding author.E-mail addresses: [email protected] (T. Czaran), [email protected]

(R.F. Hoekstra), [email protected] (D.K. Aanen).

Please cite this article in press as: Czaran, T., et al. Selection against somatic parasitism can maintain allorecognition in fungi. Fungal Genet. Biol.http://dx.doi.org/10.1016/j.fgb.2014.09.010

Tamas Czaran a, Rolf F. Hoekstra b,⇑, Duur K. Aanen b

a MTA-ELTE Research Group of Theoretical Biology and Evolutionary Ecology, Pázmány Péter sétány 1/C, 1117 Budapest, Hungaryb Laboratory of Genetics, Wageningen University, Droevendaalsesteeg 1, 6708 PB Wageningen, The Netherlands

a r t i c l e i n f o a b s t r a c t

272829303132333435363738

Article history:Received 7 July 2014Accepted 29 September 2014Available online xxxx

Keywords:CheatingFungiHeterokaryon incompatibilityKin selectionLevels of selectionSomatic incompatibility

39404142434445464748

Fusion between multicellular individuals is possible in many organisms with modular, indeterminategrowth, such as marine invertebrates and fungi. Although fusion may provide various benefits, fusionusually is restricted to close relatives by allorecognition, also called heterokaryon or somatic incompatibil-ity in fungi. A possible selective explanation for allorecognition is protection against somatic parasites.Such mutants contribute less to colony functions but more to reproduction. However, previous modelstesting this idea have failed to explain the high diversity of allorecognition alleles in nature. These modelsdid not, however, consider the possible role of spatial structure. We model the joint evolution of allorecog-nition and somatic parasitism in a multicellular organism resembling an asexual ascomycete fungus in aspatially explicit simulation. In a 1000-by-1000 grid, neighbouring individuals can fuse, but only if theyhave the same allotype. Fusion with a parasitic individual decreases the total reproductive output of thefused individuals, but the parasite compensates for this individual-level fitness reduction by a dispropor-tional share of the offspring. Allorecognition prevents the invasion of somatic parasites, and vice versa,mutation towards somatic parasitism provides the selective conditions for extensive allorecognitiondiversity. On the one hand, if allorecognition diversity did not build up fast enough, somatic parasites wentto fixation; conversely, once parasites had gone to fixation no allorecognition diversity built up. On theother hand, the mere threat of parasitism could select for high allorecognition diversity, preventinginvasion of somatic parasites. Moderate population viscosity combined with weak global dispersal wasoptimal for the joint evolution of allorecognition and protection against parasitism. Our results areconsistent with the widespread occurrence of allorecognition in fungi and the low degree of somaticparasitism. We discuss the implications of our results for allorecognition in other organism groups.

� 2014 Published by Elsevier Inc.

49

63

64

65

66

67

68

69

70

71

72

73

74

75

1. Introduction

Cooperation is predicted to evolve more easily if social interac-tions predominantly occur between genetically related individuals(Bijma and Wade, 2008; Hamilton, 1964; West et al., 2007). Posi-tive assortment between related individuals can be achieved byhigh population viscosity or by kin discrimination, which caneither be based on shared environment or on genetic cues(Grafen, 1990). Genetic, cue-dependent kin recognition is commonin all domains of life, including plants (Chen et al., 2012; Dudleyand File, 2007), fungi (Aanen et al., 2008; Glass and Dementhon,2006; Saupe et al., 2000), bacteria (Gibbs et al., 2008), vertebrates(Charpentier et al., 2007), insects (van Zweden and d’Ettorre, 2010),

76

77

78

79

slime moulds (Hirose et al., 2011; Strassmann et al., 2011) andsessile marine invertebrates (Grosberg, 1988). However, the originand maintenance of polymorphic genetic recognition cues remainincompletely understood despite substantial theoretical andempirical research (e.g. (Crozier, 1986; Nauta and Hoekstra,1994; Rousset and Roze, 2007)). In this paper, we address theevolution of a specific example of kin recognition, allorecognitionin multicellular (filamentous) fungi.

A multicellular individual essentially is a colony of cells, whichcooperate to increase their inclusive fitness, for example bydivision of labour or by size-related protection against predation(Buss, 1987; Gavrilets, 2010; Ispolatov et al., 2012; Koschwanezet al., 2011). Extant multicellular organisms represent differentstages in the transition towards individuality (Queller andStrassmann, 2009). In the most derived forms, the multicellularindividual has become the new unit of selection, as adaptationsat this level, such as an early germline–soma differentiation, render

(2014),

http://dx.doi.org/10.1016/j.fgb.2014.09.010

mailto:[email protected]




http://www.sciencedirect.com/science/journal/10871845

http://www.elsevier.com/locate/yfgbi


80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

2 T. Czaran et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx


17 October 2014

somatic cells evolutionarily dead ends (Bourke, 2011; Buss, 1987;Frank, 2003; Okasha, 2006). However, organisms with indetermi-nate growth, such as fungi, do not have an early germ–somadifferentiation, so that all body parts retain the potential to repro-duce and therefore still are ‘‘hopeful reproductives’’ (Aanen et al.,2008). Fungi further differ from other multicellular organisms intheir growth mode. They form filaments (hyphae) that are branch-ing and fusing regularly to form a dense, radially growing, networkcalled the mycelium. Each fragment can reproduce via fission orthe formation of asexual spores. In contrast to most other multicel-lular organisms, cell compartmentalization is not very strong andin some fungi nuclei can freely move through parts of the myce-lium (Roper et al., 2011). Therefore, the cooperating units in thefungal colony are the haploid nuclei (Rayner, 1991). Colony sizecan be increased through hyphal fusion between germinatingspores during colony establishment and between the hyphae ofmature colonies (Read et al., 2010; Roca et al., 2005). Fusionbetween mycelia can be mutually beneficial (Aanen et al., 2009;Bastiaans et al., submitted for publication; Pontecorvo, 1958;Richard et al., 2012).

In spite of the potential benefits of fusion between individuals,fusion between different mycelia is restricted by genetic allorecog-nition systems based on gene polymorphisms at several loci,restricting fusion almost exclusively to clonally related colonies(Aanen, 2010; Glass et al., 2000; Saupe, 2000; Saupe et al., 2000).Successful fusion between colonies requires matching at all recog-nition loci otherwise mycelia are somatically incompatible andfusion will be interrupted by programmed cell death of the fusedhyphal compartments at the border of two colonies. The wide-spread occurrence of allorecognition suggests that the disadvan-tages of fusion on average will be greater than the benefits(Nauta and Hoekstra, 1994). The most generally accepted hypoth-esis is that allorecognition has evolved to limit the opportunitiesfor somatic parasites (e.g. (Aanen et al., 2008; Buss, 1982, 1987;Buss and Green, 1985; Grafen, 1990; Grosberg and Strathmann,2007; Nauta and Hoekstra, 1994; Rousset and Roze, 2007). Asomatic parasite is a variant that contributes less to colony func-tions, but relatively more to reproduction. Within a colony of coop-erating nuclei, such a variant will be selected, but selection amongcolonies will disfavour such a variant. Thus, it is a cheater as itincreases its relative fitness within a colony of wildtype nuclei,but does so at the cost of colony fitness (Ghoul et al., 2014).Although such mutants are not common in fungi, a few examplesare known (Davis, 1960; Pittenger and Brawner, 1961).

Although it makes intuitive sense that allorecognition hasevolved as a protection against somatic parasitism, its evolutionis not well understood. First, Crozier (1986) pointed out thatshort-term selection will work against the genetic diversity of cuesrequired for allorecognition. If fusion provides a benefit, or if rejec-tion is costly, the common allele will always be favored, because itwill fuse more often than rare alleles. Therefore, allorecognition‘eats up’ the genetic variation upon which it crucially relies, a pre-diction now known as ‘Crozier’s paradox’ (Aanen et al., 2008;Crozier, 1986; Rousset and Roze, 2007). The balance betweenshort-term positive frequency-dependent selection limiting allo-recognition diversity, as predicted by Crozier, and the long-termrisk to be hit by a somatic parasite (or a ‘cheat’; (Ghoul et al.,2014; Grafen, 1990), selecting for increased allorecognition diver-sity, thus remains unknown. Second, even under the assumptionof potentially negative fitness consequences of somatic fusion, the-oretical modelling predicts only limited polymorphism for allorec-ognition, and cannot explain the extreme extent to whichpolymorphism can go in many cases (Grosberg and Quinn, 1989;Jansen and van Baalen, 2006; Nauta and Hoekstra, 1994). Althoughthe Nauta and Hoekstra model could explain the maintenance of alimited number of allotypes once they were already above a certain

Please cite this article in press as: Czaran, T., et al. Selection against somatic pahttp://dx.doi.org/10.1016/j.fgb.2014.09.010

threshold frequency in the population, it could not explain theinvasion of new allotypes starting from very low frequencies. How-ever, this model did not take spatial structure into account.

In the present study, we test the hypothesis that the potential forsomatic parasitism can select for allorecognition, and vice versa,that allorecognition keeps somatic parasites at a low frequency,using a spatially explicit model. We model the joint evolution ofallorecognition and somatic parasitism in an asexual multicellularascomycete fungus with the potential for somatic fusion. In ourmodel, initially no parasitism and no allorecognition exist, i.e. everyindividual can fuse with every other. Mutation at the parasite locuscan generate a nuclear parasite, while mutation at the allorecogni-tion locus can generate new allorecognition types (allotypes).Somatic fusion is only possible between individuals with identicalallotypes. The parasitic allele may spread horizontally by somaticfusion in compatible inter-individual confrontations. We systemat-ically assess the effect of different assumptions about the details ofparasitism and about the fitness consequences of fusion, and espe-cially about the effect of spatial structure on both the allorecogni-tion diversity and the level of parasitism evolving. Our studyshows that allorecognition contributes to the stability of multicel-lular growth by preventing the invasion of somatic parasites, andvice versa, that the potential for somatic parasitism can select forextensive allorecognition diversity, thus solving Crozier’s paradox.

2. Methods

2.1. The model

The model is a spatially explicit cellular automaton (CA) withwhich we addressed the joint evolution of allorecognition andsomatic parasitism in a modular, sedentary multicellular organism,which produces propagules, such as spores (Fig. 1). However –since application to other biological systems in which somaticfusion occurs is straightforward – we will use a more general ter-minology in the text throughout.

The basic assumptions of the model are the following:

1. Multicellular individuals are sedentary and occupy a 2Dhabitat represented by a 1000 � 1000 square lattice oftoroidal topology. Each site of the lattice harbours onemulticellular individual.

2. The organisms are haploid and reproduction is exclusivelyasexual through mitotic propagule formation. Therefore,we can simplify the genetic specification of the allorecog-nition system by assuming a single locus with a maximumof 50 different alleles. Thus the population contains 50 dif-ferent allorecognition types or allotypes at most.

3. The individuals are identical in all but two respects: theymay carry different alleles at the allorecognition locus,and either a parasitic (h) or a non-parasitic (H) allele at a‘‘Parasitism’’ locus (which can be a functionally connectedset of loci, of course). Every allorecognition allele canmutate with small probability (10�6 per generation) intoany other one of the 49. Also, a non-parasitic H individualmay produce a mutant parasitic h offspring with probabil-ity 10�6 per generation; no back mutation from parasitesto non-parasites is allowed. We have also tested a tenfoldhigher mutation rate towards parasitism (10�5 pergeneration).

4. Neighbouring individuals can fuse if they have the sameallorecognition allele. Such fusions may result in extendedchimaeric individuals that occupy more than one patch, butthe actual effects of fusion remain local – each individualcomponent in the chimaera feels the effect of fusion with

rasitism can maintain allorecognition in fungi. Fungal Genet. Biol. (2014),


207

208

209

210

211

212

213

214

215

216

217

218

219

221221

222

223

224

225

226

227

228

229

230231

232

233

234

235

236

237

238239

241241

242

243

244

245

246

247

248

249

250252252

253255255

256

257

258

259260262262

263265265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

281281

282

284284

285286

287

288

289

290

291

292

293

294

295

296

297

298

Fig. 1. Key aspects of the theoretical model and the life cycle of the modelled multicellular organism in a spatially structured environment of 1000 � 1000 patches.Reproduction is asexual and somatic fusion can occur with neighbouring individuals of the same allotype.

T. Czaran et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx 3


17 October 2014

just its four immediate neighbours (i.e., within its Neumannneighbourhood). For extended individuals that occupymore than one patch, all calculated fitness values are scaledto the propagule output of a solitary non-parasitic individ-ual in a single patch.

One possible effect of fusion is a fitness benefit (in terms ofpropagule production) depending on the number of fused neigh-bours according to a ‘‘diminishing returns’’ scheme, i.e. the moreneighbours a certain individual is fused with the higher its fitness,but the fitness gain becomes smaller with each additional individ-ual fused to the assembly:

W 0 ¼W0 þ bN þ n� 1

N þ n;

where W’ is the fitness of a focal individual or individual component(occupying one patch in the grid) after fusion, W0 is the fitness of asolitary non-parasitic individual, b is the fitness gain parameter, andN and n are the numbers of non-parasitic (‘‘H’’) and parasitic (‘‘h’’)neighbours, respectively, having the allotype of the focal individual.Note that N + n P 1, as the focal individual is always part of its ownneighbourhood, and that the fitness benefit of fusion itself does notdepend on the fused individuals being parasitic or non-parasitic.5. The other effect of fusion is a possible fitness loss due to para-

sitism: parasitic ‘‘h’’ alleles decrease the (local) fitness of theindividual they invade, and this cost of parasitism is a linearfunction of the proportion of parasitic individuals among thosemaking up the chimaera. With this assumption the fitness W(scaled per patch in the grid) of a fused unit containing Nnon-parasitic and n parasitic individuals is.

W ¼W 0 N þ 1� sð ÞnN þ n

� �¼ W0 þ b

N þ n� 1N þ n

� �N þ 1� sð Þn

N þ n

� �:

s is the fitness cost of parasitism; 0 6 s 6 1. We shall use the term‘‘Codominant’’ to label this linear algorithm of fitness phenotypedetermination, noting that this terminology is admittedly some-what sloppy, since our model organism is not a diploid one. Thisalgorithm is different from that of Nauta and Hoekstra (1994),who consider an all-or-none (step) type fitness cost function: fusedunits pay the cost of parasitism only if they consist entirely ofparasitic individuals, i.e.,


W ¼W 0 if N–0

W ¼W 0ð1� sÞ if N ¼ 0:

This scenario will be referred to as ‘‘Recessive’’. For comparisonwe also use another step function for the cost of parasitism, assum-ing that even a single parasitic individual in the fused unit is suffi-cient for the entire fitness cost to apply:

W ¼W 0 if n ¼ 0

W ¼W 0ð1� sÞ if n–0:

This will be called the ‘‘Dominant’’ algorithm. We have repeatedour simulations with all these three cost functions.

6. In a chimaeric individual containing fused parasitic and non-par-asitic parts, parasitic h alleles gain a disproportionally higherchance than non-parasitic H alleles of being transmitted duringreproduction. The measure of this segregation distortion is H,the proportion of h propagules in the yield of a chimaeric individ-ual with equal numbers of H and h alleles (H > 0.5), so that therelative abundances of non-parasitic (WH) and parasitic (Wh)propagules produced by the focal individual (given in units ofthe propagule output of a solitary non-parasitic individual) are

WH ¼Wð1�HÞN

1�Hð ÞN þHn

Wh ¼WHn

1�Hð ÞN þHn:

7. The propagules are dispersed evenly over the sites within a cer-tain distance from the mother-individual, i.e. propagule dis-persal is local and scalable. We used local dispersalneighbourhoods of sizes 3 � 3, 11 � 11 and 21 � 21 sites(D = 1, 5 and 10, respectively; D is the radius of the dispersalneighbourhood). A small fraction g of the propagules is dis-persed ‘‘globally’’: all the sites of the lattice have equal chanceto receive a globally dispersed propagule. The effect of globaldispersal was tested at g = 0%, 1%, 5% and 10%.

8. Propagule production is followed by the death of all motherindividuals – at each site the next generation is started up froma propagule, which is randomly chosen from among thosepresent at that site.



299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342



17 October 2014

Diversity was measured using the Gini–Simpson Index (Jost,2006), which is the probability that two individuals taken fromthe population at random belong to different groups.

3. Results

Starting all simulations from a single allorecognition type withonly non-parasitic genotypes, and allowing for rare (of probability10�6 per individual per generation) mutations at the allorecogni-tion locus and at the parasite locus, we ask two questions:

1. How many different allorecognition specificities doevolve depending on the various parameters (the fitnessbenefit of fusion b, the type and the parameter s of thefitness cost function, the segregation distortion H andpropagule dispersal distance D)?

2. Under what combinations of these parameters does thesystem expel the parasitic h allele, maintain H/h poly-morphism, or drive the H allele to extinction?

We have explored the biologically feasible part of the parameterspace of the model, in search for possible patterns of polymorphismregarding both allorecognition and somatic parasitism. A represen-

Fig. 2. Some typical predictions of the model. Each chart shows the change with timeparasitic (black) and non-parasitic individuals (grey) after 1000 generations. A simulationallotypes and towards somatic parasitism. Grey bands: non-parasitic strains; black bandsthe location of new allotypes does not indicate their origin, as they are randomly placed oto parasitism’’ algorithm, except for b3 (which used the ‘‘Recessive’’ scenario) and c3 (wifor all the simulations; mutation rate from non-parasitic to parasitic genotype is lP = 10non-parasitic. All other simulation parameter values are specified on the graphs. For a s


tative ‘‘walk’’ in the parameter space is given in Fig. 2. A more sys-tematic scan is given in Figs. S1–S3 in the Supplement. One of thegeneral observations is that the long-term coexistence of parasitesand non-parasites is very improbable in this model. Any long-termpolymorphism for parasitism we could find was temporal, i.e.,oscillatory, and even that was confined to a very small part of theparameter space. Wherever we could see coexistence of parasitesand non-parasites at generation 10,000 (like in the case ofFig. 2a2), longer simulations proved that coexistence was transitory– either parasites or non-parasites take over sooner or later. Anotherconspicuous feature of the results is that if the non-parasitic H alleledefeats the parasitic h allele, it does so thanks to the selection of newallorecognition mutations, sometimes many of them, resulting in ahigh allotype diversity (Fig. 2). In some of the cases when the para-site takes over, the final state of the population is still polymorphicfor allorecognition (like in Fig. 2b1). This might be considered as theremnant of past, failed attempts of non-parasitic strains to disposeof their parasitic mutants.

We will first consider the consequences of applying the threedifferent cost-of-parasitism algorithms and of changing keyparameters of the model before we explore the effect of spatialstructure on the coevolution between allotype diversity andsomatic parasitism.

in the number and frequency of allotypes and the extent to which these consist ofstarts with a single, non-parasitic allotype, where mutation can occur towards new

: parasitic strains. Different allotypes are separated with white lines. Please note thatn the vertical axis. All the graphs are produced with the ‘‘Codominant fitness loss dueth the ‘‘Dominant’’ algorithm). Mutation rate at the allorecognition locus: lA = 10�6

�6 everywhere except for a3 where lP = 10�5. Parasitic cells do not mutate back toystematic scan of the parameter space see the Supplement Figs. S1–S3.



343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466



17 October 2014

3.1. The shape of the cost-of-parasitism function

The shape of the cost-of-parasitism function has a clear effect onthe coevolution of allorecognition and somatic parasitism: the‘‘Recessive’’ step function (which represents the weakest selectionagainst parasitism) yields high parasite frequencies and a low num-ber of allotypes – in fact no new allotypes evolve and all cellsbecome parasitic in most of the ‘‘Recessive’’ cases, at least if fusionitself carries a fitness benefit (Fig. 2b3; S1). Linear (‘‘Codominant’’)fitness loss due to parasitism results in a high diversity of allotypes(Fig. 2b2; S2), but only at high segregation asymmetry (H = 0.8) andmostly at stronger selection against parasitism (s = 0.2). At low seg-regation asymmetry (H = 0.6) and relatively intensive spatial mix-ing (D P 5) the system tends to oscillate, with only two allotypespresent in the lattice at any point of time (Fig. 2c2), which seemsto be a structurally unstable property of the model: even smallchanges in parameter values annihilate these oscillations. Thediversity pattern of allorecognition types obtained with the ‘‘Dom-inant’’ scenario (step-like fitness-loss function dropping fitness toits minimum with a single parasitic individual in the fused unit) isqualitatively similar to that of the ‘‘Codominant’’ algorithm(Fig. 2c3; S3), except that no oscillations like on Fig. 2c2 show up.Relatively strong selection (s = 0.2) against parasitized units is anecessary condition for allotype diversity to build up, just like inthe other two scenarios.

3.2. Segregation distortion (H)

Segregation distortion has an interesting threefold effect onallorecognition evolution. Apart from the trivial tendency of para-sites to be more likely to exclude non-parasites from the steadystate of the simulations if H is larger, another, less obvious effectis that allorecognition diversity increases with H in all the caseswhere parasites are finally abolished (e.g., compare Fig. 2c2–b2,and panels A to B in Figs. S1–S3). At weaker counter-parasite selec-tion (s = 0.1), increased segregation distortion always allows theparasite to take over, but the population might become persis-tently polymorphic on the allorecognition locus during the exclu-sion process of non-parasitic individuals (Fig. 2b1).

The surprising increase in the number of allorecognition typeswith increasing the segregation distortion H can be interpretedas an evolutionary reaction to the increased selection pressurefrom parasitic alleles: new non-parasitic allorecognition mutantsenjoy the advantage of not immediately being parasitized, so theycan spread as long as their population is small and therefore notlikely to quickly produce its own parasitic mutant. That is, the pop-ulations of small and ‘‘clean’’ (parasite-free) allotypes persist andincrease, and the larger the parasitic pressure (i.e., the larger H)is, the higher their initial advantage will be relative to other, para-sitized populations.

The third effect of increasing segregation distortion is that thefrequency distribution of allotypes become more even. The morecommon an allotype becomes, the higher its chance to be hit bya parasitic mutant. This leads to frequency-dependent selectionfavouring rare allotypes, thus resulting in a frequency distributionof the mutant allotypes tending close to even (Nauta and Hoekstra,1994). The larger H is, the more pronounced this frequency-equal-izing effect seems to be, probably due to the consequent increasedselection pressure from parasitism.

3.3. The fitness benefit of fusion (b)

We have compared the cases of zero (b = 0.0) and 10% (b = 0.1)fitness advantage of fusion. The most conspicuous differencebetween these two parameter values was that, in the absence ofa fusion benefit (b = 0.0), the chance of parasitic nuclei to take over


was lower (Fig. 2a1and a3). That is, the advantage of parasitismincreases with the fitness benefit of fusion (compare the left twocolumns to the right two columns of Figs. S1–S3). This may seemsurprising at first, given that the fusion benefit is aspecific: itapplies to the same extent to parasites and non-parasites. Thisresult can be understood by considering that parasitic invasioncan only occur through fusions; therefore, higher fitness rewardsfor fusion add an extra fitness bonus to parasites. This result canalso be understood from the perspective of non-parasitic allotypes.As new, non-parasitic, allorecognition groups will initially be at alow frequency, they will have few partners to fuse with. If fusionis beneficial, this low starting frequency will hamper their inva-sion. As allotype diversity reduces the opportunities for parasites,lower allotype diversity increases the opportunities for parasites.

The advantage of parasitism due to the aspecific fitness rewardof fusion can go so far that, in a considerable section of the param-eter space, parasitic alleles become fixed at b = 0.1, in contrast tothe corresponding simulations with b = 0.0 in which fusion is notbeneficial in itself, and non-parasites win the game by evolvingmany allotypes (this can be clearly seen in Fig. S2B, for example).

3.4. The local dispersal of propagules (D)

The local dispersal of propagules (D) has a principal effect diag-onally opposite to most previous model conclusions regarding par-asitism (Frank, 1996). Conventional wisdom says that the parasitealways does better if it can disperse efficiently within the host pop-ulation(Hamilton and May, 1977). In contrast to this expectation,the great majority of our simulations show that it is the non-para-sitic strain that performs better with increasing D. The general pat-tern seems to be that increasing dispersal increases the number ofnew allotypes invading the system, which then may or may not befollowed by non-parasite takeover (compare Fig. 2a2–b2). A feasi-ble explanation may be based on the fact that the fate of an allorec-ognition type depends on its ability to – at least temporarily –escape from its own parasitic mutants. Parasitic mutants originatefrom non-parasites of the same allotype, and they can gain extra fit-ness by fusing with other individuals of their ‘‘mother strain’’, thusharvesting the fitness benefits due to fusion and also to segregationdistortion. This can be achieved easily if dispersal does not drift theparasitic mutant far away from its mother strain. In fact dispersingfar away is most probably fatal for new parasites, because with avery high probability they lose both sources of fitness gain (thosefrom b and from H). Viewed from the non-parasites’ aspect, dis-persal is the best escape for a relatively rare allotype from its ownparasitic mutants; it can keep the allotype free of parasites for along time. If the cost of parasitism is low (s = 0.1) and segregationdistortion is strong (H = 0.8), the ultimate winner can still be theparasite. Also under these conditions, the more advantage wasgiven to non-parasitic allotype mutants by dispersal at the begin-ning, the more parasitic allorecognition groups will be present inthe stationary phase of the process (compare within-column panelsof Figs. S1–S3). That is, allotype diversity is correlated to dispersal,even if parasites win the game finally.

The spatial patterns generated by fusion dynamics and limitedpropagule dispersal are almost always patchy to some extent. Thisis a direct consequence of two different effects: the aspecific ben-efit of fusion (wherever it is assumed) which selects for allotypesstaying aggregated, and the crucial dependence of parasiticmutants on their ability to aggregate with their own hosts. Thestatistical effects of these forces on the distribution of local fusionsare summarized in Fig. 3.

3.5. The global dispersal of propagules (g)

The global dispersal of propagules (g) has proven to be of atwofold effect on allotype diversity: it may help or hinder the



467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

Fig. 3. The distribution of the 5-site local fusion units within the spatial patterns of the 10,000th generations. Horizontal axis: number of parasitic individuals fused to thefocal individual; Vertical axis: Number of non-parasitic individuals fused to the focal individual. (The focal individual itself is also counted.) Darker colours represent higherfrequencies; the stars show the average parasite/non-parasite compositions of local fusion units. (For interpretation of the references to color in this figure legend, the readeris referred to the web version of this article.)



17 October 2014

evolution of new allotypes. Fig. 4 shows the effect of increasingglobal spore dispersal in the ‘‘Codominant’’ (probably the mostrealistic) series of simulations, for both segregation asymmetriesstudied (H = 0.6 and 0.8). It is obvious that for very high segrega-tion asymmetry (H = 0.8) global dispersal is always deleterious –it helps parasites take over (Fig. 4a). For moderate asymmetry(H = 0.6), however, weak global dispersal (at g = 1–5%) isbeneficial, but strong global dispersion (g > 5–10%) is also delete-rious (Fig. 4b).

493

494

495

496

497

498

499

500

501

502

503

504

4. Discussion

Somatic fusion provides opportunities for somatic parasites.The overall conclusion of our study is that allorecognition can bean efficient means to prevent the spread of somatic parasites andvice versa, that the potential of parasitism can select for extensiveallorecognition diversity. Thus, selection against somatic parasit-ism provides a solution to Crozier’s paradox and explains theextensive allorecognition found in populations of fungal speciesand other organisms with modular growth.


4.1. A comparison with previous models on the coevolution ofcooperation and allorecognition

In contrast to most other studies, our model assumes coopera-tion, viz. among the cells of a multicellular individual, as a startingpoint, and then considers the effect of cue-dependent fusionamong individuals on the opportunities for somatic parasites.Other theoretical studies have sought to explain the evolution ofcue-dependent cooperation bottom-up, i.e. starting with solitarybehaviour (e.g. (Axelrod et al., 2004; Czaran and Hoekstra, 2009;Jansen and van Baalen, 2006; Lee et al., 2012; Rousset and Roze,2007). Multicellular growth followed by possible fusion amongmulticellular individuals contains two layers of cooperation, i.e.altruistic cooperation among cells within the individual and mutu-ally beneficial cooperation between fusing individuals. In ourmodel, ‘defecting’ means ‘parasitizing’ because we assume a trade-off between somatic parasitism and somatic growth (which weshow is reasonable). Therefore, in contrast to previous models(Axelrod et al., 2004; Czaran and Hoekstra, 2009; Jansen and vanBaalen, 2006; Lee et al., 2012; Rousset and Roze, 2007), a parasitehas reduced basic fitness when it grows solitarily.



505

506

507

508

509

510

511

512

Fig. 4. The effect of global dispersal on the evolution and the maintenance of allotype diversity. All simulations were run with the ‘‘Codominant’’ algorithm. Fixed parameters:relative fitness benefit of fusion (b = 0.1) and relative fitness loss due to parasitism (s = 0.2) (a) for the case of segregation asymmetry H = 0.8; (b) for H = 0.6. Grey bands: non-parasitic strains; black bands: parasitic strains.



17 October 2014

Because of the theoretical problems to explain the maintenanceof allorecognition diversity, it has been argued that allotypediversity could also be maintained if recognition loci have asecondary function that is subject to diversifying selection


(Crozier, 1986; Rousset and Roze, 2007). For example, if allorecog-nition loci also function in recognition of parasites (of a differentspecies), they may be under diversifying selection because ofcoevolution with these (Paoletti and Saupe, 2009). Another



513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638



17 October 2014

possibility is disassortative mating based on allotypic differences,and a recent paper shows that this can stabilise allotype diversity(Holman et al., 2013). Although these hypotheses are not incom-patible with the results of this paper, our results demonstrate thatit is not necessary to infer secondary functions of allorecognitionlabels, although further research is needed to study the assump-tions and parameter values of our model.

4.2. The benefit of fusion

We confirm Crozier’s (Crozier, 1986) prediction that, if fusion isbeneficial, short-term selection works against allorecognitiondiversity. This in turn increases the probability that the parasitetakes over. Thus, selection in favour of fusion means selectionagainst the invasion of new allorecognition types at the same time,because new allotypes prevent fusions and thus also the harvest ofthe benefit thereof. Since the chances of non-parasitic genotypes toexclude parasitic invasion depend on their ability to hitch-hike onnew allorecognition mutants, any selective force that prevents newallotypes from spreading is advantageous for the parasite.

But is fusion between different individuals beneficial? The factthat in many groups of organisms somatic fusion, either betweenindividuals or between tissues, is a physiological possibility doessuggest that it is. Possible advantages of inter-individual fusionare genetic complementation, mitotic recombination or moreeffective resource utilization. Also fusion between genetically iden-tical individuals has been shown to be advantageous, presumablydue to more efficient reproduction with increased size (Aanenet al., 2009; Amar et al., 2008; Bastiaans et al., submitted forpublication; Raymundo and Maypa, 2004; Richard et al., 2012).The benefit of fusion between individuals probably depends on size(Aanen et al., 2008; Amar et al., 2008). Experiments show thatfusion is beneficial if individuals are small, but not if they are large(Bastiaans et al., submitted for publication). In our model, no ben-efit of fusion (b = 0) implies that a multicellular individual withinone patch is large enough to yield the full benefit of multicellular-ity, so that additional increase in size because of between-individ-ual fusion is neutral.

Alternatively, fusion between cells or hyphae within a single(physical) individual provides an advantage (Nauta and Hoekstra,1994). Fusion among fragmented individuals of the same clonewould then be an artefact of this intra-organismal fusion and hencea consequence of an imperfect non-self-recognition, rather than akin-recognition system (Aanen et al., 2008). For fungi, the signifi-cance of intra-individual fusion has indeed been shown (Rayner,1991; Rayner et al., 1984; Xiang et al., 2002). Also for colonial mar-ine invertebrates, intra-individual fusion is probably important(Feldgarden and Yund, 1992; Hughes et al., 2004).

4.3. The role of dispersal

Previous models have shown that the evolution of allorecogni-tion is difficult to understand in the absence of population struc-ture(Grosberg and Quinn, 1989; Hartl et al., 1975; Nauta andHoekstra, 1994; Rousset and Roze, 2007; Tsutsui, 2004). Here weanalysed the effect of population viscosity, using three levels oflocal dispersion, and, for a subset, local dispersal in combinationwith varying levels of global dispersion. The relationship betweendispersion and the evolution of allorecognition and parasitism wascomplex. Dispersal changes the balance between selection at twohierarchical levels, i.e. among nuclei within individuals, and amongindividuals. In general, some dispersal favours among-individualselection relative to within-individual selection, and thus favoursallorecognition diversity, and disfavours parasitism. Furthermore,if fusion provides a benefit, dispersal affects the efficiency ofpositive frequency-dependent selection of allotypes.


In the simulations without any global dispersion, generally themost efficient local dispersion (D = 10) was optimal for the build-up of allorecognition diversity and selection against somatic para-sitism (compare Figs. 2b2-2a2). This can be understood as follows.A new, mutant allotype free of parasites, freely spreads in the hab-itat until it produces its own parasite by mutation. With low dis-persal, this parasitic mutant will remain close to its non-parasitichost allotype, so it has a good chance to exploit it. With more dis-persal, this chance decreases. In addition, with long-distance dis-persal the host individuals are more dispersed (they do not formcontiguous patches), so that even if the parasite gets in contactwith a few hosts, it will not have access to the whole host popula-tion. Once it has excluded all the hosts locally available, the para-site has a disadvantage compared to non-parasitic, individuals ofother allotypes, with which it cannot fuse, and this is why it goesextinct.

The effect of some global dispersal upon the evolution of allo-recognition diversity and somatic parasitism is subtle. On the onehand, global dispersal allows parasites to spread efficientlythrough the population, limiting the scope for parasite-free indi-viduals to evolve a new allotype. A high percentage of global dis-persion (g P 10%) indeed always led to fixation of parasitism inthe absence of any allotype diversity (Fig. 4). On the other hand,some global dispersal helps new non-parasitic allotype mutantsfind and invade parasitic patches of other allotypes, i.e., sites inwhich they have a fitness advantage compared to their parasiticneighbours.

Not much is known on the dispersal rates of relevant organismsin nature. Grosberg and Quinn (1986) found that in Botryllus schlos-seri allorecognition types tend to cluster, more than if larval dis-persal was random. This result has been confirmed in laterstudies, although rare dispersal further away also occurs (Ben-Shlomo et al., 2008; Grosberg, 1987). However, actual dispersalmay have been underestimated in these studies, because cleansubstrates were available close to a source colony. If nearby sub-strates are covered, larvae are forced to swim further, so that theactual dispersal rate will be higher than measured in these exper-iments (Grosberg, 1987). Social bacteria have been found to havestrong population structure, even at small spatial scales, demon-strating that dispersal is limited (Vos and Velicer, 2008). In fungi,spores are generally wind-dispersed, which is highly efficient. Nev-ertheless, it has been shown that the great majority of spores willnot disperse far (<100 m) from their origin (Lacey, 1996;Wolfenbarger, 1946). A recent study on spore dispersal in six ecto-mycorrhiza-forming basidiomycetes even showed that 95% of bas-idiospores fall within 58 cm of the cap (Galante et al., 2011).

4.4. The nature of parasitism

A few studies have addressed somatic parasitism in multicellu-lar organisms. In the ascomycete fungus Neurospora crassa, whichresembles the organism modelled in this paper, Pittenger andBrawner (1961) described a mutant that increased in frequencywithin the mycelium without an obvious effect on mycelium fit-ness, so that it is not a parasite as defined here. However, usingan auxotrophic marker, this mutant artificially became a parasite.Under selective conditions, the mutant could only grow as a chi-maera in combination with the wild-type nuclei due to the auxot-rophy, but within the chimaera, the parasite increased in frequencyup to the point where mycelial growth ceased. Davis (Davis, 1960)described a different parasite. From a pantothenate-requiringmutant (Pan�), he isolated a mutant that could grow on low con-centrations of pantothenate (Pan+). A chimaera of these two geno-types could also grow on this low concentration, but duringgrowth, the Pan� nuclei took over within the mycelium, reducingmycelial fitness up to the point where growth stopped. In these



639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727Q4

728

729

730

731Q5Q6

732

733

734

735

736

737

738739740741742Q7743744745746747748749Q8750751752753754755756757758759760761762763764765



17 October 2014

two studies, the effect of the parasites thus was most consistentwith a linear cost-of-parasitism function, although in a slightlymore complex fashion than modelled here.

De Boer (1995)) modelled the evolution of allorecognition as ameans to protect ‘the genetic integrity’ of an individual. He arguedthat restriction of somatic fusion may function to preserve specificevolved adaptations. A particular individual adapted to oneenvironment, may be maladapted to a different environment, andanother individual vice versa. If these individuals fuse, themaladapted individual will be a parasite relative to the well-adapted genotype, and these roles will be reverted in the otherenvironment. Somatic parasitism, therefore, is a relative and notan absolute characteristic of an individual, depending on the envi-ronment and the individual with which it fuses. The examples onfungi above show that this indeed may be the case, as the effectof an auxotrophic mutation will depend on the fusing partner,and on the environment, i.e. on the presence of the nutrient. Cru-cially, these examples also show that for some mutants the costssaved on somatic functions, i.e. the uptake or production of nutri-ents, can be invested in personal reproductive success, makingsuch mutants genuine parasites as modelled in this paper.

In Botryllus, heritable differences have been found betweengenotypes in fused individuals for replacement of germline andsomatic cells (De Boer, 1995; Laird et al., 2005; Stoner et al.,1999; Stoner and Weissman, 1996). Note, however, that the mod-elled organism resembles a modular organism without a germ-soma differentiation. Therefore, we do not distinguish betweensomatic and germline parasites, as has been done for the colonialprotochordate B. schlosseri (Stoner et al., 1999).

Recently, Kuzdzal-Fick et al. (2011) experimentally selected forsomatic parasitism under conditions of low relatedness amongcells in Dictyostelium, where multicellular fruiting bodies form byaggregation of cells. In nature, relatedness among cells in this spe-cies is high via a combination of kin recognition and efficient dis-persal of spores (Kuzdzal-Fick et al., 2011; Ostrowski et al.,2008). However, experimentally manipulated conditions of lowrelatedness favoured the evolution of obligate parasites, whichcould not produce fruiting bodies alone, but only in combinationwith non-parasites.

Not only nuclear parasitic genes can lower fitness of the multi-cellular individual, but also cytoplasmic parasites. For example, infungi, suppressive mitochondria (Bertrand, 2000) and deleteriousplasmids and viruses are widespread (Ghabrial and Suzuki,2009). Such selfish genetic elements can infect other individualsfollowing somatic fusion. Therefore, allorecognition not only canprotect against infection with nuclear parasites (Debets andGriffiths, 1998), but potentially also against infection withcytoplasmic parasites (Brusini et al., 2011). Several studies haveindeed found that allorecognition provides some protectionagainst infection with cytoplasmic elements, although not perfect(Debets et al., 1994). Our model could also be regarded as coveringthe case of a parasitic cytoplasmic element, by assuming that hindividuals represent cytoplasmically infected individuals. Raremutations to novel parasitic nuclei could then be interpreted asrare infections between individuals of different allorecognitiontype. To simulate that a fraction of the offspring of infected individ-uals can be cured, especially after sexual reproduction, we wouldneed to extend our model to allow for mutations to an H genotypein a propagule from an h individual.

4.5. Conclusion

Our study shows that allorecognition can stabilise multicellu-larity in organisms with somatic fusion and vice versa, that thethreat of somatic parasitism provides the selective conditions forthe maintenance of allorecognition, thus solving Crozier’s paradox.


Under many reasonable combinations of parameter values, weobtained a high degree of allorecognition, in combination with avery low degree, or absence, of parasitism. If fusion provides a ben-efit, the evolution of allorecognition requires more stringentconditions, confirming Crozier’s (Crozier, 1986) prediction. Popula-tion viscosity with some long-range dispersal generally was mostfavourable for the evolution of allorecognition diversity andselection against parasitism.

Idiosyncrasies of specific organisms will determine the poten-tial for somatic parasitism and the need for allorecognition. Futurestudies therefore need to address basic details of multicellulargrowth, somatic fusion and the scale of competition in differentspecies under natural conditions. Furthermore, experimental stud-ies are needed to measure the mutation rate towards parasitismand the characteristics of somatic parasites. For example, severalloss-of-function mutations in fungi have been found to be somaticparasites in combination with wild-type genotypes (Davis, 1960).If somatic parasitism is a general characteristic of loss-of-functionmutants, this increases the potential number of mutations causingsomatic parasitism dramatically. Also the effect of parasiticmutants upon individual fitness, and whether this is ‘dominant’or ‘recessive’ will need to be determined. This knowledge is neces-sary to validate our theoretical predictions for specific cases.

5. Uncited reference

Wilson and Grosberg (2004).

Acknowledgments

The authors thank Eric Bastiaans and Fons Debets for useful dis-cussion, and Alan Grafen and Niels Anten for comments on themanuscript. DKA was funded by a grant of NWO (vidi) and TC byvisitor’s grants provided by NWO, PE&RC and OTKA Grant No.K100806.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.fgb.2014.09.010.

References

Aanen, D.K. et al., 2008. The social evolution of somatic fusion. Bioessays 30, 1193–1203.

Aanen, D.K. et al., 2009. High symbiont relatedness stabilizes mutualisticcooperation in fungus-growing termites. Science 326, 1103–1106.

Aanen, D.K., Debets, A.J.M., Glass, N.L., Saupe, S.J., 2010. Biology and genetics ofvegetative incompatibility in fungi. In: Borkovich, K., Ebbole, D.J., (Eds.), Cellularand Molecular Biology of Filamentous Fungi.

Amar, K.-O. et al., 2008. Coral kin aggregations exhibit mixed allogenic reactionsand enhanced fitness during early ontogeny. BMC Evol. Biol. 8.

Axelrod, R. et al., 2004. Altruism via kin-selection strategies that rely on arbitrarytags with which they coevolve. Evolution 58, 1833–1838.

Bastiaans, E., et al., submitted for publication. Experimental demonstration of thebenefits of somatic fusion and the consequences for allorecognition. Evolution.

Ben-Shlomo, R. et al., 2008. Pattern of settlement and natural chimerism in thecolonial urochordate Botryllus schlosseri. Genetica 132, 51–58.

Bertrand, H., 2000. Role of mitochondrial DNA in the senescence and hypovirulenceof fungi and potential for plant disease control. Ann. Rev. Phytopathol. 38, 397–422.

Bijma, P., Wade, M.J., 2008. The joint effects of kin, multilevel selection and indirectgenetic effects on response to genetic selection. J. Evol. Biol. 21, 1175–1188.

Bourke, A.F.G., 2011. Principles of Social Evolution. Oxford University Press, Oxford.Brusini, J. et al., 2011. Parasitism and maintenance of diversity in a fungal vegetative

incompatibility system: the role of selection by deleterious cytoplasmicelements. Ecol. Lett. 14, 444–452.

Buss, L.W., 1982. Somatic cell parasitism and the evolution of somatic tissuecompatibility. Proc. Natl. Acad. Sci. USA 79, 5337–5341.

Buss, L.W., 1987. The Evolution of Individuality. Princeton University Press,Princeton.




766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835

836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905



17 October 2014

Buss, L.W., Green, D.R., 1985. Histoincompatibility in vertebrates: the relicthypothesis. Dev. Comp. Immunol. 9, 191–201.

Charpentier, M.J.E. et al., 2007. Kin discrimination in juvenile mandrills, Mandrillussphinx. Anim. Behav. 73, 37–45.

Chen, B.J.W. et al., 2012. Detect thy neighbor: identity recognition at the root levelin plants. Plant Sci. 195, 157–167.

Crozier, R.H., 1986. Genetic clonal recognition abilities in marine invertebrates mustbe maintained by selection for something else. Evolution 40, 1100–1101.

Czaran, T., Hoekstra, R.F., 2009. Microbial communication, cooperation andcheating: quorum sensing drives the evolution of cooperation in bacteria.PLoS ONE 4.

Davis, R.H., 1960. Adaptation in pantothenate-requiring Neurospora. 2. Nuclearcompetition during adaptation. Am. J. Bot. 47, 648–654.

De Boer, R.J., 1995. The evolution of polymorphic compatibility molecules. Mol. Biol.Evol. 12, 494–502.

Debets, A.J.M., Griffiths, A.J.F., 1998. Polymorphism of het-genes prevents resourceplundering in Neurospora crassa. Mycol. Res. 102, 1343–1349.

Debets, F. et al., 1994. Vegetative incompatibility in Neurospora – its effect onhorizontal transfer of mitochondrial plasmids and senescence in naturalpopulations. Curr. Gen. 26, 113–119.

Dudley, S.A., File, A.L., 2007. Kin recognition in an annual plant. Biol. Lett. 3, 435–438.

Feldgarden, M., Yund, P.O., 1992. Allorecognition in colonial marine invertebrates –does selection favor fusion with kin or fusion with self? Biol. Bull. 182, 155–158.

Frank, S.A., 1996. Models of parasite virulence. Quart. Rev. Biol. 71, 37–78.Frank, S.A., 2003. Perspective: repression of competition and the evolution of

cooperation. Evolution 57, 693–705.Galante, T.E. et al., 2011. 95% of basidiospores fall within 1 m of the cap: a field- and

modeling-based study. Mycologia 103, 1175–1183.Gavrilets, S., 2010. Rapid transition towards the division of labor via evolution of

developmental plasticity. PLOS Comput. Biol. 6.Ghabrial, S.A., Suzuki, N., 2009. Viruses of plant pathogenic fungi. Annu. Rev.

Phytopathol., 353–384.Ghoul, M. et al., 2014. Toward an evolutionary definition of cheating. Evol. Int. J.

Org. Evol. 68, 318–331.Gibbs, K.A. et al., 2008. Genetic determinants of self identity and social recognition

in bacteria. Science 321, 256–259.Glass, N.L., Dementhon, K., 2006. Non-self recognition and programmed cell death

in filamentous fungi. Curr. Opin. Microbiol. 9, 553–558.Glass, N.L. et al., 2000. The genetics of hyphal fusion and vegetative incompatibility

in filamentous ascomycete fungi. Annu. Rev. Genet. 34, 165–186.Grafen, A., 1990. Do animals really recognize kin? Anim. Behav. 39, 42–54.Grosberg, R.K., 1987. Limited dispersal and proximity-dependant mating success in

the colonial ascidian Botryllus schlosseri. Evolution 41, 372–384.Grosberg, R.K., 1988. The evolution of allorecognition specificity in clonal

invertebrates. Quart. Rev. Biol. 63, 377–412.Grosberg, R.K., Quinn, J.F., 1986. The genetic control of kin recognition by the larvae

of a colonial marine invertebrate. Nature 322, 456–459.Grosberg, R.K., Quinn, J.F., 1989. The evolution of selective aggression conditioned

on allorecognition specificity. Evolution 43, 504–515.Grosberg, R.K., Strathmann, R.R., 2007. The evolution of multicellularity: a minor

major transition? Ann. Rev. Ecol., Evol., Syst. 38, 621–654.Hamilton, W.D., 1964. Genetical evolution of social behaviour I+II. J. Theor. Biol. 7,

1–34.Hamilton, W.D., May, R.M., 1977. Dispersal in stable habitats. Nature 269, 578–581.Hartl, D.L. et al., 1975. Adaptive significance of Vegetative Incompatibility in

Neurospora crassa. Genetics 81, 553–569.Hirose, S. et al., 2011. Self-recognition in social amoebae is mediated by allelic pairs

of tiger genes. Science 333, 467–470.Holman, L. et al., 2013. Crozier’s paradox revisited: maintenance of genetic

recognition systems by disassortative mating. BMC Evol. Biol. 13, pp. 211–211.Hughes, R.N. et al., 2004. Kin or self-recognition? Colonial fusibility of the bryozoan

Celleporella hyalina. Evol. Dev. 6, 431–437.Ispolatov, I. et al., 2012. Division of labour and the evolution of multicellularity.

Proc. Roy. Soc. B-Biol. Sci. 279, 1768–1776.Jansen, V.A.A., van Baalen, M., 2006. Altruism through beard chromodynamics.

Nature 440, 663–666.Jost, L., 2006. Entropy and diversity. Oikos 113, 363–375.Koschwanez, J.H. et al., 2011. Sucrose utilization in budding yeast as a model for the

origin of undifferentiated multicellularity. PLOS Biol. 9.

906


Kuzdzal-Fick, J.J. et al., 2011. High relatedness is necessary and sufficient tomaintain multicellularity in Dictyostelium. Science 334, 1548–1551.

Lacey, J., 1996. Spore dispersal – its role in ecology and disease: the Britishcontribution to fungal aerobiology. Mycol. Res. 100, 641–660.

Laird, D.J. et al., 2005. Stem cells are units of natural selection in a colonial ascidian.Cell 123, 1351–1360.

Lee, W. et al., 2012. An evolutionary mechanism for diversity in siderophore-producing bacteria. Ecol. Lett. 15, 119–125.

Nauta, M.J., Hoekstra, R.F., 1994. Evolution of vegetative incompatibility infilamentous ascomycetes. 1. Deterministic models. Evolution 48, 979–995.

Okasha, S., 2006. Evolution and the Levels of Selection. Oxford University Press,Oxford.

Ostrowski, E.A. et al., 2008. Kin discrimination increases with genetic distance in asocial amoeba. PLOS Biol. 6, 2376–2382.

Paoletti, M., Saupe, S.J., 2009. Fungal incompatibility: evolutionary origin inpathogen defense? Bioessays 31, 1201–1210.

Pittenger, T., Brawner, T.G., 1961. Genetic control of nuclear selection in Neurosporaheterokaryons. Genetics 46, 1645-&.

Pontecorvo, G., 1958. Trends in Genetic Analysis New York. Columbia UniversityPress, p. 145.

Queller, D.C., Strassmann, J.E., 2009. Beyond society: the evolution of organismality.Philos. Trans. Roy. Soc. B-Biol. Sci. 364, 3143–3155.

Raymundo, L.J., Maypa, A.P., 2004. Getting bigger faster: mediation of size-specificmortality via fusion in juvenile coral transplants. Ecol. Appl. 14, 281–295.

Rayner, A.D.M., 1991. The challenge of the individualistic mycelium. Mycologia 83,48–71.

Rayner, A.D.M. et al., 1984. The biological consequences of the individualisticmycelium. In: Jennings, D.H., Rayner, A.D.H. (Eds.), The Ecology and Physiologyof the Fungal Mycelium. Cambridge University Press, Cambridge, pp. 509–540.

Read, N., et al., 2010. Hyphal Fusion. In: Borkovich, K.A.D.E., (Ed.), Cellular andMolecular Biology of Filamentous Fungi American Society of Microbiology. pp.260–273.

Richard, F. et al., 2012. Cooperation among germinating spores facilitates thegrowth of the fungus, Neurospora crassa. Biol. Lett. 8, 419–422.

Roca, M.G. et al., 2005. Conidial anastomosis tubes in filamentous fungi. FemsMicrobiol. Lett. 249, 191–198.

Roper, M. et al., 2011. Nuclear and genome dynamics in multinucleate ascomycetefungi. Curr. Biol. 21, R786–R793.

Rousset, F., Roze, D., 2007. Constraints on the origin and maintenance of genetic kinrecognition. Evolution 61, 2320–2330.

Saupe, S.J., 2000. Molecular genetics of heterokaryon incompatibility in filamentousascomycetes. Microbiol. Mol. Biol. Rev. 64, 489–502.

Saupe, S.J. et al., 2000. Vegetative incompatibility in filamentous fungi: Podosporaand Neurospora provide some clues. Curr. Opin. Microbiol. 3, 608–612.

Stoner, D.S., Weissman, I.L., 1996. Somatic and germ cell parasitism in a colonialascidian: possible role for a highly polymorphic allorecognition system. Proc.Natl. Acad. Sci. USA 93, 15254–15259.

Stoner, D.S. et al., 1999. Heritable germ and somatic cell lineage competitions inchimeric colonial protochordates. Proc. Natl. Acad. Sci. USA 96, 9148–9153.

Strassmann, J.E. et al., 2011. Kin discrimination and cooperation in microbes. Ann.Rev. Microbiol. 65 (65), 349–367.

Tsutsui, N.D., 2004. Scents of self: the expression component of self/nonselfrecognition systems. Ann. Zool. Fenn. 41, 713–727.

van Zweden, J., d’Ettorre, P., 2010. Nestmate recognition in social insects and therole of hydrocarbons. In: AG, B., GJ, B. (Eds.), Insect Hydrocarbons: Biology,Biochemistry and Chemical Ecology. Cambridge University Press, Cambridge,pp. 222–243.

Vos, M., Velicer, G.J., 2008. Isolation by distance in the spore-forming soil bacteriumMyxococcus xanthus. Curr. Biol. 18, 386–391.

West, S.A. et al., 2007. Evolutionary explanations for cooperation. Curr. Biol. 17,R661–R672.

Wilson, A.C.C., Grosberg, R.K., 2004. Ontogenetic shifts in fusion-rejectionthresholds in a colonial marine hydrozoan, Hydractinia symbiolongicarpus.Behav. Ecol. Sociobiol. 57, 40–49.

Wolfenbarger, D.O., 1946. Dispersion of small organisms – distance dispersion ratesof bacteria, spores, seeds, pollen and insects – incidence rates of disease andinjuries. Am. Midland Nat. 35, 1–152.

Xiang, Q.J. et al., 2002. The ham-2 locus, encoding a putative transmembraneprotein, is required for hyphal fusion in Neurospora crassa. Genetics 160, 169–180.



fungal genetics and biologyreal.mtak.hu/21277/1/selection against somatic parasitism...1 3 selection...

Documents